AI ALIGNMENT FORUMTags
AF

Iterated Amplification

EditHistorySubscribe
Discussion (0)
Help improve this page (1 flag)
EditHistorySubscribe
Discussion (0)
Help improve this page (1 flag)
Iterated Amplification
Random Tag
Contributors
2jacobjacob
2Ben Pace

Iterated Amplification is an approach to AI alignment, spearheaded by Paul Christiano. In this setup, we build powerful, aligned ML systems through a process of initially building weak aligned AIs, and recursively using each new AI to build a slightly smarter and still aligned AI. 

See also: Factored cognition. 

Posts tagged Iterated Amplification
Most Relevant
4
13Iterated Distillation and Amplification
Ajeya Cotra
4y
7
3
41Paul's research agenda FAQ
Alex Zhu
5y
34
3
42Challenges to Christiano’s capability amplification proposal
Eliezer Yudkowsky
5y
2
3
27A guide to Iterated Amplification & Debate
Rafael Harth
2y
0
2
11AlphaGo Zero and capability amplification
Paul Christiano
4y
16
2
62Debate update: Obfuscated arguments problem
Beth Barnes
2y
14
1
43My Understanding of Paul Christiano's Iterated Amplification AI Safety Research Agenda
Chi Nguyen
2y
9
1
67An overview of 11 proposals for building safe advanced AI
Evan Hubinger
3y
25
1
38My Overview of the AI Alignment Landscape: A Bird's Eye View
Neel Nanda
1y
4
1
48Writeup: Progress on AI Safety via Debate
Beth Barnes, Paul Christiano
3y
15
1
29Relaxed adversarial training for inner alignment
Evan Hubinger
3y
15
1
21Prize for probable problems
Paul Christiano
5y
0
1
23Garrabrant and Shah on human modeling in AGI
Rob Bensinger
1y
7
1
22Corrigibility
Paul Christiano
4y
2
1
16Factored Cognition
Andreas Stuhlmüller
4y
1
Load More (15/47)
Add Posts